The Dollar Is Eroding. Nothing Is Replacing It. — The Full Research Note

Ask why the dollar is losing ground and you’ll usually get a name in reply: gold, the yuan, Bitcoin. Something, the story goes, is quietly taking the dollar’s place. Twenty-five years of central bank reserve data tell a stranger story: the decline is real, but nothing is stepping into the gap it leaves behind.

The U.S. dollar’s share of global central bank reserves has fallen from roughly 71% in 1999 to 57.1% in the first quarter of 2026, according to the International Monetary Fund’s Currency Composition of Official Foreign Exchange Reserves survey — the most complete public record of how the world’s central banks actually hold their money. That decline isn’t a rounding error, and it isn’t just the dollar’s exchange rate wobbling around from one quarter to the next. Multiple independent studies, including work published by the Federal Reserve itself, have tested and confirmed that it reflects real, deliberate portfolio decisions by reserve managers. This note works through where that money has gone, and why the three usual suspects — gold, the yuan, crypto — each turn out to explain only a sliver of it.

01 — FramingThe Slow Bleed

Reserve currency status isn’t a switch a currency flips on or off. It’s a share, measured continuously, and the IMF has tracked it in a consistent way since 1999 through its COFER survey. In that first year, the dollar accounted for roughly 71% of the foreign exchange reserves disclosed by the roughly 150 reporting central banks. By early 2026, that figure had fallen to 57.1%.

The path down was never a straight line, and it’s worth resisting the temptation to read it as one. The decline was noticeably faster between 2015 and 2019 than in the years immediately before or after, and the dollar’s share actually ticked up in the most recent quarter on record — from 56.4% at the end of 2025 to 57.1% in the first quarter of 2026. The IMF’s own account of that uptick is a useful reminder of how much these quarterly moves owe to currency valuation rather than active diversification: roughly half of the increase, the Fund noted, was simply the dollar appreciating against other reserve currencies, not central banks buying more of it. Strip out those valuation swings more broadly, as researchers at the Federal Reserve Bank of New York have done, and the underlying trend is still downward — just less dramatic than the raw headline number on its own suggests.

What the COFER data can’t tell you on its own is whether this reflects a broad, coordinated move away from the dollar, or the outsized actions of a handful of large reserve holders rebalancing their own books. That distinction matters enormously for what comes next, and it’s the question the next two sections work through in turn.

02 — MechanismWhy Reserve Status Is Self-Reinforcing

Reserve currency status, the depth of a government bond market, and a currency’s share of global trading and payments aren’t three separate facts about the dollar. They’re three measurements of one underlying loop, and each leg reinforces the others. Central banks hold dollar reserves largely because U.S. Treasuries are the deepest, most liquid safe asset available anywhere in the world. Treasuries are that liquid substantially because of the constant buying pressure created by reserve managers, exporters settling trade in dollars, and financial institutions that need dollar collateral to operate. Weaken any one leg and the other two feel it; strengthen any one leg and the other two benefit. There’s no single point at which the loop was “caused” — it behaves like a feedback system, not a chain of dominoes.

The scale of that loop today is hard to overstate. The Bank for International Settlements’ 2025 Triennial Central Bank Survey — the most comprehensive official measure of global currency trading, run once every three years since 1986 — found that foreign exchange markets now turn over roughly $9.6 trillion a day, a record, up 28% from 2022, driven in part by the volatility surrounding 2025’s tariff announcements. The dollar sat on one side of 89% of that turnover in April 2025, actually up slightly from 88% in 2022. Away from the trading floor, the Federal Reserve’s own research finds the dollar has dominated trade invoicing everywhere outside Europe for decades: averaged over 1999 to 2019, the last period for which comprehensive invoicing data exists, roughly 96% of export invoicing in the Americas, 74% in Asia-Pacific, and 79% across the rest of the world was priced in dollars. Only within Europe, where the euro is dominant at 66%, did that pattern break — and nothing in the more current trading and payments data below suggests that picture has shifted meaningfully since.

Cross-border payments tell a similar, if slightly muddier, story. The Federal Reserve’s own analysis of SWIFT payment-messaging data puts the dollar’s share of international payments at roughly 50%, a figure that has ticked up slightly in recent years. Add back in payments made entirely within the eurozone, in euros, and the dollar’s share is actually higher still, near 60% — a reminder that these payment-share figures depend heavily on exactly which slice of global payment traffic gets counted, and the answer doesn’t always move in the direction intuition suggests. The lesson isn’t that any one of these numbers is the “true” figure. It’s that the dollar’s dominance looks very different depending on which slice of the plumbing you’re measuring — and reserve share, the number that gets the headlines, is consistently the slice showing the most erosion.

That gap between reserve managers moving faster than private market participants is itself informative. If the dollar’s role were unwinding as one coordinated retreat, you’d expect trading, invoicing, and reserve holdings to move together. They haven’t. Central banks, who answer to finance ministries and sometimes to explicit diversification mandates, appear to be adjusting faster than the exporters, banks, and traders who have to invoice, settle, and hedge in whatever currency is most liquid on any given day. That’s the first clue that this is a diversification story rather than a replacement story.

03 — Myth-BustingIt’s Not Going Where You Think

Here the popular narrative runs into its biggest problem. The most rigorous academic treatment of this question is a 2022 International Monetary Fund working paper by Serkan Arslanalp, Barry Eichengreen, and Chima Simpson-Bell, titled “The Stealth Erosion of Dollar Dominance.” The paper set out to test a specific alternative explanation: that the dollar’s declining reserve share since 2000 was a statistical illusion, produced by exchange-rate and interest-rate swings or by the outsized behavior of a handful of central banks with unusual balance sheets, rather than by real diversification. Their conclusion ruled that out. The decline is real, active diversification, and it’s broad-based rather than concentrated in one or two large holders.

What the paper found next rarely makes it into the popular version of this story. The decline in the dollar’s share hasn’t been matched by a corresponding rise in the shares of the other long-standing reserve currencies — the euro, the yen, and the pound sterling, which together with the dollar have historically comprised the bulk of official reserves and the IMF’s own Special Drawing Rights basket. Those currencies’ shares have stayed roughly flat over the same period. Instead, the shift out of dollars has gone in two directions: roughly a quarter of it into the Chinese renminbi, and the remaining three-quarters into a long tail of what the authors call “nontraditional” reserve currencies — the Australian dollar, the Canadian dollar, the Swiss franc, and other currencies that have historically played only a minor role as reserve assets.

That’s close to the opposite of what an actual successor-currency story ought to look like. If reserve managers believed the yuan, or any single currency, was positioned to take over the dollar’s role, you’d expect the diversification flow to concentrate there. It hasn’t. It has diffused into a dozen or more currencies, individually too small, too shallow, or too capital-controlled to plausibly anchor the global financial system on their own. The renminbi itself remains capped near 2% of global reserves, according to the IMF’s most recent figures, held back chiefly by China’s continued capital controls and the absence of a fully open domestic bond market that foreign central banks can rely on. The most accurate one-line description of what’s happening to the dollar’s reserve share isn’t “replacement.” It’s diffusion.

04 — The CaveatGold Buying Isn’t a Dollar Verdict

Readers of our earlier note on gold’s rally will recognize the next complication. Central bank gold buying has been genuinely extraordinary since 2022, and it’s tempting to read that buying as a direct anti-dollar statement — central banks fleeing dollar assets for a neutral one instead. A September 2025 Federal Reserve International Finance Discussion Paper by economist Colin Weiss, which examined exactly how governments have managed the mix of gold and dollar reserves over recent decades, found that this framing doesn’t hold up for most buyers. Gold reserve accumulation, the paper concludes, is generally not associated with de-dollarization of a country’s broader reserve portfolio at the individual-country level — with a handful of prominent exceptions, chiefly Russia, China, and Turkey.

A separate study quantifying the split between gold price appreciation and actual new buying makes a related point: for China, roughly 91% of the increase in gold’s share of its reserves since 2015 is explained by the rising price of gold itself, not by new purchases; for Russia, the comparable figure is about 78%. None of this means the buying isn’t happening — the People’s Bank of China has now extended its gold-buying streak to 21 consecutive months as of July 2026, the longest run on record since it began publishing monthly figures in 2015 — but it does mean the “central banks are fleeing the dollar for gold” framing overstates what most individual reserve managers are actually doing. For the typical central bank, gold buying looks a great deal like ordinary portfolio diversification that happens to coincide with an extraordinary run in the gold price, not a deliberate statement against the dollar.

05 — HistoryThe Handover Nobody Remembers Correctly

Precision matters here, because the only real precedent for an actual reserve currency handover — sterling losing its position to the dollar — is far stranger than the tidy version usually told. The conventional account holds that the pound sterling remained the dominant reserve currency until after the Second World War, when an exhausted Britain finally ceded the position to an ascendant United States. For decades, that was the textbook story.

It’s also, according to Barry Eichengreen’s own archival research with economic historian Marc Flandreau, largely wrong. Working from newly assembled data on the currency composition of central bank reserves in the 1920s and 1930s, Eichengreen and Flandreau found that the dollar first overtook sterling as the leading reserve currency far earlier — around 1924, more than two decades before the date assumed by earlier scholars. The two currencies had run roughly neck and neck for much of the early 1920s, with the dollar pulling ahead as New York’s financial markets matured and the newly created Federal Reserve System began actively supporting a market in dollar-denominated trade credit. By the second half of the 1920s, sterling was, in the authors’ words, continuously “under the harrow.”

Then it got messier still. Sterling actually reclaimed the lead in the early 1930s, after the U.S. devalued the dollar in 1933 — a reversal the conventional story leaves out entirely, because it doesn’t fit a narrative of smooth, one-directional decline. It took the disruptions of the Great Depression, Britain’s departure from the gold standard, and ultimately the Second World War itself before the dollar’s advantage became durable, sometime in the early 1940s, permanently this time, and well before Bretton Woods formalized the arrangement in 1944. Eichengreen and Flandreau state their own conclusion plainly: the advantages of incumbency in the competition for reserve currency status are considerably weaker than commonly supposed — and dominance, once lost, is not necessarily gone for good.

Read that way, the real precedent doesn’t neatly support either the “it takes decades” framing or the “it can happen overnight” framing. It supports something closer to: reserve currency leadership can change hands faster than incumbency advantages suggest, but a change in leadership and a permanent, uncontested handover are two different events, and the gap between them was measured in nearly two full decades of a real, two-way contest, punctuated by at least one clear reversal, before the Second World War settled the matter for good.

06 — InstitutionsStablecoins Export the Dollar

Cryptocurrency is the third candidate regularly proposed as the dollar’s eventual successor, and it repays a closer look, because the actual data runs directly against the popular framing. The Federal Reserve’s own 2025 research note on the dollar’s international role puts a precise figure on it: approximately 99% of stablecoin market capitalization — the digital tokens designed to hold a constant value, which function as the working medium of exchange within crypto markets — is linked to the U.S. dollar. Tether and USD Coin, the two largest stablecoins by a wide margin, are both dollar-pegged.

That composition matters because of what stablecoins are used for in practice: settling crypto trades, and increasingly, providing a form of dollar-denominated savings and payments access in countries with unstable local currencies or limited access to the formal U.S. banking system. In both cases, the effect is to extend the reach of the dollar into markets and users that traditional dollar banking infrastructure can’t easily serve, not to create a true alternative to it. Washington has, if anything, leaned into that dynamic rather than fought it: the GENIUS Act, signed into law on July 18, 2025, built the first federal regulatory framework specifically for dollar-backed stablecoins, requiring full reserve backing and regular audits — formal government infrastructure built around the assumption that dollar-linked stablecoins are here to stay. The Fed’s own researchers put the broader implication bluntly: increased stablecoin usage may cause more emerging-market economies to become effectively dollarized, not less. A technology built by some of its early proponents to route around the traditional financial system has, so far, mostly reproduced the dollar’s dominance inside that new system rather than displacing it.

07 — The CatalystSanctions as Accelerant

The clearest recent catalyst for active reserve diversification isn’t in dispute, and we’ve covered it at length elsewhere on Trimline Research: the February 2022 freezing of roughly $300 billion of Russia’s foreign currency reserves by the United States and its allies, in response to the invasion of Ukraine. That single event demonstrated to every other reserve manager on earth, in real time, that a reserve asset held abroad and denominated in a foreign currency depends entirely on that foreign country’s continued political cooperation. It’s the single most plausible explanation for the acceleration in both gold buying and broader reserve diversification since 2022, and it sits underneath both this note and our earlier work on gold’s rally.

08 — The MilestoneTwo Trend Lines, Not Yet Crossed

Two real trends are running in parallel right now, and they deserve to be held side by side without overstating what their coexistence proves. The first is China’s central bank buying streak, which the World Gold Council put at 21 consecutive months as of July 2026 — the longest run since the People’s Bank of China began publishing monthly figures in 2015 — with official holdings climbing to roughly 2,366 tonnes and the pace of purchases actually accelerating through the summer even as gold traded well below the record high it set in January 2026. This is the same buying wave that, as we detailed in Rethinking Gold’s Rally, coincided with gold’s reported overtaking of U.S. Treasuries as the world’s largest reserve asset in 2025, a milestone the European Central Bank itself flagged in its June 2026 report on the euro’s international role.

That comparison deserves a caveat, and a timely one: a Federal Reserve research note published just days before this piece went to press argues the “gold overtook Treasuries” framing is more misleading than it looks. Most of the shift, Weiss finds, reflects gold’s price surge rather than a jump in central bank buying — that surge was driven mainly by private investors piling into gold-backed ETFs from late 2024 onward, not by official reserve managers, whose purchases had merely held at their already-elevated post-2022 pace. It’s also skewed by a handful of countries — the U.S., Germany, Italy, France, and the IMF — sitting on legacy gold holdings acquired before the Bretton Woods system ended in 1971, which together account for roughly half of all gold reserves worldwide and haven’t meaningfully added to them since. Strip out valuation effects and those legacy holders, and foreign official Treasury holdings were still running some $600 billion ahead of official gold holdings as of mid-2026; by dollar value, Treasuries even reclaimed the outright lead over non-U.S. gold reserves by June 2026. The IMF’s own COFER team reaches a strikingly similar conclusion in its most recent data brief, noting plainly that gold’s overtaking of Treasuries in 2025 “was driven almost entirely by gold price valuation effects” and isn’t reflected at all in the dollar’s own COFER share, which has stayed broadly stable throughout. The reserve-composition headlines aren’t wrong, exactly — they’re just measuring something narrower than they appear to. It’s the same lesson this note keeps returning to: a striking cross-asset comparison and an actual shift in central bank behavior are not always the same thing.

The second real trend is the U.S. fiscal trajectory: debt held by the public stood at 99% of GDP at the end of fiscal year 2025, and the Congressional Budget Office’s own current-law baseline, published in its Budget and Economic Outlook: 2026 to 2036, projects that figure climbing to 120% of GDP by 2036, a level higher than at any point in the nation’s history, surpassing even the post-World War II peak.

Neither trend, on its own, tells you the dollar’s reserve role is approaching a crisis point. Debt-to-GDP ratios and reserve currency status have historically been more loosely correlated than intuition suggests — Japan has run debt-to-GDP ratios above 200% for years without any comparable erosion of the yen’s, admittedly much smaller, international role. What the two trend lines together actually support is a more modest and more defensible claim: the pressures pushing reserve managers to diversify away from the dollar are structural, and they are not obviously fading, even as no single alternative has yet emerged capable of absorbing that diversification at scale.

09 — The VisionThree Doors, None Open Yet

The natural next question isn’t just what isn’t working — it’s what a real transition would require, and whether anything happening right now is actually moving toward it. Three doors get proposed most often. On the present evidence, none of them are open.

The renminbi door. Genuine reserve-currency status requires actual capital-account convertibility — letting money flow freely in and out of a country, rather than largely at the state’s discretion, so foreign central banks can hold and transact in the currency the way they hold dollars or euros today. A 2024 study by Barry Eichengreen and coauthors that tested this question directly found convertibility isn’t just helpful but close to a precondition for real reserve-currency status, and Beijing has given no indication it intends to loosen capital controls in that direction. That’s the direct explanation for why the renminbi’s reserve share, having briefly touched almost 3% in 2021, has since drifted back down toward 2% — despite active Chinese efforts including expanded central bank swap lines and Belt and Road renminbi lending, according to Arslanalp, Eichengreen, and Simpson-Bell’s most recent data.

The BRICS door. Talk of a common BRICS currency, or a wholesale shift to settling trade in local currencies to sidestep the dollar, has circulated since at least the bloc’s 2023 and 2024 summits. The same three researchers’ own read on where that stands, published in March 2026: these discussions “have failed to gain any momentum.” A common currency requires the kind of fiscal, monetary, and political coordination among member states that the eurozone spent decades building among far more aligned economies to begin with, and the BRICS bloc — which now spans Brazil, Russia, India, China, South Africa, and a growing list of newer members with sharply divergent interests — is not obviously any closer to that.

The digital-currency door. Central bank digital currencies were once framed as the technology that could finally let cross-border payments route around dollar-clearing infrastructure entirely. The most advanced pilot, Project mBridge — linking the central banks of China, Hong Kong, Thailand, the UAE, and Saudi Arabia — lost its neutral overseer when the Bank for International Settlements withdrew from active involvement in October 2024, after which China’s own digital yuan came to account for the large majority of the platform’s settlement volume, by industry estimates. Rather than producing one interoperable rail the whole world could use instead of the dollar system, the more likely outcome looks like two competing blocs: mBridge, increasingly shaped by Chinese participation, and a rival Western-anchored platform, Project Agorá, built around the Federal Reserve Bank of New York alongside central banks in Europe, Japan, Korea, and Mexico. If that’s where cross-border digital-currency infrastructure ends up, the result isn’t a single successor currency — it’s a more fragmented payments system with the dollar still anchoring the larger of the two resulting blocs.

None of these three amounts to anything close to a live transition. What this note’s evidence does support is a fourth, far less dramatic possibility: durable multipolarity, in which no single currency ever fully replaces the dollar, and the erosion documented throughout this piece simply continues at its present grinding pace, diffusing further into whichever assets and currencies happen to look safe and liquid in a given year. Arslanalp, Eichengreen, and Simpson-Bell — whose research this note has leaned on throughout — put it almost exactly that way in their most recent update: the only forecast that can be made with confidence is that the international reserve system will continue to evolve. That’s an unsatisfying answer to “what comes next,” but on the evidence assembled here, it may be the honest one.

The one scenario worth watching, according to the same researchers, isn’t a rival currency at all — it’s the possibility that the dollar’s own safe-haven status could weaken: the property that makes investors and central banks want to hold dollars specifically because they can be trusted to hold their value when everything else is falling apart. That got a live test in April 2025, when the announcement of “reciprocal” tariffs triggered a bout of market stress in which the dollar fell alongside risk assets rather than rising the way a safe-haven currency normally would during a shock. The episode proved temporary; the dollar’s safe-haven characteristics reasserted themselves within weeks. But it’s the one crack in the feedback loop described in Section 02 that, if it widened and became persistent rather than temporary, would matter far more than any single rival currency — because it would undermine the very reason reserve managers hold dollars in the first place.